bartz 0.2.0__py3-none-any.whl → 0.3.0__py3-none-any.whl

bartz/prepcovars.py

@@ -10,10 +10,10 @@
 # to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 # copies of the Software, and to permit persons to whom the Software is
 # furnished to do so, subject to the following conditions:
-# 
+#
 # The above copyright notice and this permission notice shall be included in all
 # copies or substantial portions of the Software.
-# 
+#
 # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 # IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 # FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
@@ -55,10 +55,10 @@ def quantilized_splits_from_matrix(X, max_bins):
     """
     out_length = min(max_bins, X.shape[1]) - 1
     # return _quantilized_splits_from_matrix(X, out_length)
-    @functools.partial(jaxext.autobatch, max_io_nbytes=500_000_000)
-    def func(X):
+    @functools.partial(jaxext.autobatch, max_io_nbytes=2 ** 29)
+    def quantilize(X):
         return _quantilized_splits_from_matrix(X, out_length)
-    return func(X)
+    return quantilize(X)
 
 @functools.partial(jax.vmap, in_axes=(0, None))
 def _quantilized_splits_from_matrix(x, out_length):
@@ -82,8 +82,36 @@ def _quantilized_splits_from_matrix(x, out_length):
     max_split = max_split.astype(jaxext.minimal_unsigned_dtype(out_length))
     return splits, max_split
 
-@jax.jit
-def bin_predictors(X, splits):
+@functools.partial(jax.jit, static_argnums=(1,))
+def uniform_splits_from_matrix(X, num_bins):
+    """
+    Make an evenly spaced binning grid.
+
+    Parameters
+    ----------
+    X : array (p, n)
+        A matrix with `p` predictors and `n` observations.
+    num_bins : int
+        The number of bins to produce.
+
+    Returns
+    -------
+    splits : array (p, num_bins - 1)
+        A matrix containing, for each predictor, the boundaries between bins.
+        The excluded endpoints are the minimum and maximum value in each row of
+        `X`.
+    max_split : array (p,)
+        The number of cutpoints in each row of `splits`, i.e., ``num_bins - 1``.
+    """
+    low = jnp.min(X, axis=1)
+    high = jnp.max(X, axis=1)
+    splits = jnp.linspace(low, high, num_bins + 1, axis=1)[:, 1:-1]
+    assert splits.shape == (X.shape[0], num_bins - 1)
+    max_split = jnp.full(*splits.shape, jaxext.minimal_unsigned_dtype(num_bins - 1))
+    return splits, max_split
+
+@functools.partial(jax.jit, static_argnames=('method',))
+def bin_predictors(X, splits, **kw):
     """
     Bin the predictors according to the given splits.
 
@@ -98,6 +126,8 @@ def bin_predictors(X, splits):
         `m` is the maximum number of splits; each row may have shorter
         actual length, marked by padding unused locations at the end of the
         row with the maximum value allowed by the type.
+    **kw : dict
+        Additional arguments are passed to `jax.numpy.searchsorted`.
 
     Returns
     -------
@@ -105,9 +135,9 @@ def bin_predictors(X, splits):
         A matrix with `p` predictors and `n` observations, where each predictor
         has been replaced by the index of the bin it falls into.
     """
-    return _bin_predictors(X, splits)
-
-@jax.vmap
-def _bin_predictors(x, splits):
-    dtype = jaxext.minimal_unsigned_dtype(splits.size)
-    return jnp.searchsorted(splits, x).astype(dtype)
+    @functools.partial(jaxext.autobatch, max_io_nbytes=2 ** 29)
+    @jax.vmap
+    def bin_predictors(x, splits):
+        dtype = jaxext.minimal_unsigned_dtype(splits.size)
+        return jnp.searchsorted(splits, x, **kw).astype(dtype)
+    return bin_predictors(X, splits)

bartz-0.3.0.dist-info/METADATA

@@ -0,0 +1,77 @@
+Metadata-Version: 2.1
+Name: bartz
+Version: 0.3.0
+Summary: A JAX implementation of BART
+Home-page: https://github.com/Gattocrucco/bartz
+License: MIT
+Author: Giacomo Petrillo
+Author-email: info@giacomopetrillo.com
+Requires-Python: >=3.10,<4.0
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Requires-Dist: jax (>=0.4.23,<0.5.0)
+Requires-Dist: jaxlib (>=0.4.23,<0.5.0)
+Requires-Dist: numpy (>=1.25.2,<2.0.0)
+Requires-Dist: scipy (>=1.11.4,<2.0.0)
+Project-URL: Bug Tracker, https://github.com/Gattocrucco/bartz/issues
+Project-URL: Repository, https://github.com/Gattocrucco/bartz
+Description-Content-Type: text/markdown
+
+[![PyPI](https://img.shields.io/pypi/v/bartz)](https://pypi.org/project/bartz/)
+
+# BART vectoriZed
+
+A branchless vectorized implementation of Bayesian Additive Regression Trees (BART) in JAX.
+
+BART is a nonparametric Bayesian regression technique. Given predictors $X$ and responses $y$, BART finds a function to predict $y$ given $X$. The result of the inference is a sample of possible functions, representing the uncertainty over the determination of the function.
+
+This Python module provides an implementation of BART that runs on GPU, to process large datasets faster. It is also good on CPU. Most other implementations of BART are for R, and run on CPU only.
+
+On CPU, bartz runs at the speed of dbarts (the fastest implementation I know of), but using half the memory. On GPU, the speed premium depends on sample size; with 50000 datapoints and 5000 trees, on an Nvidia Tesla V100 GPU it's 12 times faster than an Apple M1 CPU, and this factor is linearly proportional to the number of datapoints.
+
+## Links
+
+- [Documentation (latest release)](https://gattocrucco.github.io/bartz/docs)
+- [Documentation (development version)](https://gattocrucco.github.io/bartz/docs-dev)
+- [Repository](https://github.com/Gattocrucco/bartz)
+- [Code coverage](https://gattocrucco.github.io/bartz/coverage)
+
+## Other BART packages
+
+- [stochtree](https://github.com/StochasticTree) C++ library with R and Python bindings taylored to researchers who want to make their own BART variants
+- [bnptools](https://github.com/rsparapa/bnptools) Feature-rich R packages for BART and some variants
+- [dbarts](https://github.com/vdorie/dbarts) Fast R package
+- [bartMachine](https://github.com/kapelner/bartMachine) Fast R package, supports missing predictors imputation
+- [SoftBART](https://github.com/theodds/SoftBART) R package with a smooth version of BART
+- [bcf](https://github.com/jaredsmurray/bcf) R package for a version of BART for causal inference
+- [flexBART](https://github.com/skdeshpande91/flexBART) Fast R package, supports categorical predictors
+- [flexBCF](https://github.com/skdeshpande91/flexBCF) R package, version of bcf optimized for large datasets
+- [XBART](https://github.com/JingyuHe/XBART) R/Python package, XBART is a faster variant of BART
+- [BART](https://github.com/JingyuHe/BART) R package, BART warm-started with XBART
+- [XBCF](https://github.com/socket778/XBCF)
+- [BayesTree](https://cran.r-project.org/package=BayesTree) R package, original BART implementation
+- [bartCause](https://github.com/vdorie/bartCause) R package, pre-made BART-based workflows for causal inference
+- [stan4bart](https://github.com/vdorie/stan4bart)
+- [VCBART](https://github.com/skdeshpande91/VCBART)
+- [monbart](https://github.com/jaredsmurray/monbart)
+- [mBART](https://github.com/remcc/mBART_shlib)
+- [SequentialBART](https://github.com/mjdaniels/SequentialBART)
+- [sparseBART](https://github.com/cspanbauer/sparseBART)
+- [pymc-bart](https://github.com/pymc-devs/pymc-bart)
+- [semibart](https://github.com/zeldow/semibart)
+- [CSP-BART](https://github.com/ebprado/CSP-BART)
+- [AMBARTI](https://github.com/ebprado/AMBARTI)
+- [MOTR-BART](https://github.com/ebprado/MOTR-BART)
+- [bcfbma](https://github.com/EoghanONeill/bcfbma)
+- [bartBMAnew](https://github.com/EoghanONeill/bartBMAnew)
+- [BART-BMA](https://github.com/BelindaHernandez/BART-BMA) (superseded by bartBMAnew)
+- [gpbart](https://github.com/MateusMaiaDS/gpbart)
+- [GPBART](https://github.com/nchenderson/GPBART)
+- [bartpy](https://github.com/JakeColtman/bartpy)
+- [BayesTreePrior](https://github.com/AlexiaJM/BayesTreePrior)
+- [BayesTree.jl](https://github.com/mathcg/BayesTree.jl)
+- [longbet](https://github.com/google/longbet)
+

bartz-0.3.0.dist-info/RECORD

@@ -0,0 +1,13 @@
+bartz/BART.py,sha256=CbGzFWtYw5u38Z9-Hy3CbDXpKOOvPFAAkSqu2HZl8no,16862
+bartz/__init__.py,sha256=E96vsP0bZ8brejpZmEmRoXuMsUdinO_B_SKUUl1rLsg,1448
+bartz/_version.py,sha256=3wVEs2QD_7OcTlD97cZdCeizd2hUbJJ0GeIO8wQIjrk,22
+bartz/debug.py,sha256=9ZH-JfwZVu5OPhHBEyXQHAU5H9KIu1vxLK7yNv4m4Ew,5314
+bartz/grove.py,sha256=x_6NK_l-hrXfy1PhssYNJkX41-w_WqjDziww0E7YRS8,8500
+bartz/jaxext.py,sha256=RcVWTCGS8lXF7GBsNbKrpuA4MTcokItq0CpWm3s7CGk,12033
+bartz/mcmcloop.py,sha256=lKDszvniNXka99X3e9RCrTgvEAZHA7ZbVXEgxUYvKMY,7634
+bartz/mcmcstep.py,sha256=HPcxfl5f-OESZul-iurn0JmOnUJBe6IYTVaATeR6YBA,54221
+bartz/prepcovars.py,sha256=mMgfL-LGJ_8QpOL6iy7yfkL8A7FrT7Zfn5M3voyNwSQ,5818
+bartz-0.3.0.dist-info/LICENSE,sha256=heuIJZQK9IexJYC-fYHoLUrgj8HG8yS3G072EvKh-94,1073
+bartz-0.3.0.dist-info/METADATA,sha256=ymZNoowDdqQFyAJdeKKj6t7h8_eBXQr2cVPglyoYLDQ,4500
+bartz-0.3.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+bartz-0.3.0.dist-info/RECORD,,

bartz-0.2.0.dist-info/METADATA

@@ -1,32 +0,0 @@
-Metadata-Version: 2.1
-Name: bartz
-Version: 0.2.0
-Summary: A JAX implementation of BART
-Home-page: https://github.com/Gattocrucco/bartz
-License: MIT
-Author: Giacomo Petrillo
-Author-email: info@giacomopetrillo.com
-Requires-Python: >=3.10,<4.0
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.10
-Classifier: Programming Language :: Python :: 3.11
-Classifier: Programming Language :: Python :: 3.12
-Requires-Dist: jax (>=0.4.23,<0.5.0)
-Requires-Dist: jaxlib (>=0.4.23,<0.5.0)
-Requires-Dist: numpy (>=1.25.2,<2.0.0)
-Requires-Dist: scipy (>=1.11.4,<2.0.0)
-Project-URL: Bug Tracker, https://github.com/Gattocrucco/bartz/issues
-Project-URL: Repository, https://github.com/Gattocrucco/bartz
-Description-Content-Type: text/markdown
-
-[![PyPI](https://img.shields.io/pypi/v/bartz)](https://pypi.org/project/bartz/)
-
-# BART vectoriZed
-
-A branchless vectorized implementation of Bayesian Additive Regression Trees (BART) in JAX.
-
-BART is a nonparametric Bayesian regression technique. Given predictors $X$ and responses $y$, BART finds a function to predict $y$ given $X$. The result of the inference is a sample of possible functions, representing the uncertainty over the determination of the function.
-
-This Python module provides an implementation of BART that runs on GPU, to process large datasets faster. It is also a good on CPU. Most other implementations of BART are for R, and run on CPU only.
-

bartz-0.2.0.dist-info/RECORD

@@ -1,13 +0,0 @@
-bartz/BART.py,sha256=pRG7mALenknX2JHqY-VyhO9-evDgEC6hWBp4jpecBdM,15801
-bartz/__init__.py,sha256=E96vsP0bZ8brejpZmEmRoXuMsUdinO_B_SKUUl1rLsg,1448
-bartz/_version.py,sha256=FVHPBGkfhbQDi_z3v0PiKJrXXqXOx0vGW_1VaqNJi7U,22
-bartz/debug.py,sha256=9ZH-JfwZVu5OPhHBEyXQHAU5H9KIu1vxLK7yNv4m4Ew,5314
-bartz/grove.py,sha256=Wj_7jHl9w3uwuVdH4hoeXowimGpdRE2lGIzr4aDkzsI,8291
-bartz/jaxext.py,sha256=VYA41D5F7DYcAAVtkcZtEN927HxQGOOQM-uGsgr2CPc,10996
-bartz/mcmcloop.py,sha256=lheLrjVxmlyQzc_92zeNsFhdkrhEWQEjoAWFbVzknnw,7701
-bartz/mcmcstep.py,sha256=3ba94hXBW4UAZ11SFshnwJAgn6bpIqSZdRy_wQjEkrk,39278
-bartz/prepcovars.py,sha256=iiQ0WjSj4--l5DgPW626Qg2SSB6ljnaaUsBz_A8kFrI,4634
-bartz-0.2.0.dist-info/LICENSE,sha256=heuIJZQK9IexJYC-fYHoLUrgj8HG8yS3G072EvKh-94,1073
-bartz-0.2.0.dist-info/METADATA,sha256=LiYjTAzgoxUM2MAuaKtf0VW-_zciTKBkTX5B7HNvUbI,1490
-bartz-0.2.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-bartz-0.2.0.dist-info/RECORD,,